Adjustment of Global Adjustment of Global Gridded Precipitation Gridded Precipitation for Orographic Effects for Orographic Effects Jennifer Adam
Jan 05, 2016
Adjustment of Global Gridded Adjustment of Global Gridded Precipitation for Orographic Precipitation for Orographic
EffectsEffects
Jennifer Adam
OutlineOutline
1. Background
2. Approach
3. Application over North America
Global Gridded PrecipitationGlobal Gridded Precipitation
• Spatial Interpolation of Gauge Measurements over Land Areas
The Orographic Effect on The Orographic Effect on PrecipitationPrecipitation
Interpolation from Valley GaugesInterpolation from Valley Gauges
*
PRISMPRISM((PParameter-elevation arameter-elevation RRegressions on egressions on
IIndependent ndependent SSlopes lopes MModel)odel)
• 2.5 minute• Topographic
facets• Regresses P
against elevation on each facet
ApproachApproach
1. Select Correction Domain and “Slope Bands”
2. Select Set of Basins that overlap with Correction Domain (must be gauged)
3. Determine “Actual” Basin Average Precipitation using Sankarasubramanian (2002)
4. Determine Scaling Ratios for each “Slope Band”
5. Apply the Scaling Ratios to an Existing Gridded Precipitation Dataset
Outline of StepsOutline of Steps
Select Correction DomainSelect Correction Domain
1. Identified according to slope:
- slopes calculated from 5-minute DEM
- aggregated to half-degree
2. Set Slope Threshold
- 6 m/km (somewhat arbitrary)
3. Break Correction Domain into Slope Bands
- six bands in correction domain
SlopeSlope
> 6m/km
> 12m/km
> 18m/km
> 24m/km
> 30m/km
> 36m/km
Correction DomainCorrection Domain
1. Select Correction Domain and “Slope Bands”
2. Select Set of Basins that overlap with Correction Domain (must be gauged)
3. Determine “Actual” Basin Average Precipitation using Sankarasubramanian (2002)
4. Determine Scaling Ratios for each “Slope Band”
5. Apply the Scaling Ratios to an Existing Gridded Precipitation Dataset
Outline of StepsOutline of Steps
World Streamflow Gauge World Streamflow Gauge StationsStations
1. Select Correction Domain and “Slope Bands”
2. Select Set of Basins that overlap with Correction Domain (must be gauged)
3. Determine “Actual” Basin Average Precipitation using Sankarasubramanian (2002)
4. Determine Scaling Ratios for each “Slope Band”
5. Apply the Scaling Ratios to an Existing Gridded Precipitation Dataset
Outline of StepsOutline of Steps
“In mountainous areas, the best precipitation maps are derived by distributing streamflow back on the watershed and correcting for
evapotranspiration.”-Harry W. Anderson
Water BalanceWater Balance
GQEPdt
dS
QEP
In General:
Long term mean over watershed:
“Q” obtained from streamflow measurements
Problem: how to estimate “E”
1. Use VIC Model to Estimate “E”- relies on good precipitation estimates
- relies on empirical parameters
2. Relate “E” to Potential Evapotranspiration (PET)- no need to rely on precipitation
Alternatives for Estimating Alternatives for Estimating EvapotranspirationEvapotranspiration
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 0.5 1 1.5 2 2.5 3PET/P
E/P
Budyko (1974) CurveBudyko (1974) Curve
EnergyLimited
MoistureLimited
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 0.5 1 1.5 2 2.5 3PET/P
E/P
Budyko (1974) CurveBudyko (1974) Curve
EnergyLimited
MoistureLimited
P
PETf
P
E
1. Semi-empirical relationship
2. Independent of energy and water balance equations.
3. Mean discrepancy between the E/P ratio calculated from curves and that derived by water balance amounts to 10% (Budyko and Zubenok, 1961).
4. Applies to river basins of “considerable” size – runoff dictated by climatic factors
5. Are other variables important?
Discussion of Budyko CurveDiscussion of Budyko Curve
1. Milly (1994):
- soil plant-available water holding capacity, various seasonality parameters
2. Zhang et al. (2001): - plant-available water coefficient, w (2.0 for
forests, 0.5 for pasture and up to 1.0 for mixed vegetation)
3. Sankarasubrumanian and Vogel (2002):
- soil moisture storage capacity
Improvements to Budyko CurveImprovements to Budyko Curve
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 1 2 3PET/P
E/P
budyko gamma=1.5gamma=1.0 gamma=0.5
Sankarasubramanian et al. (2002) Sankarasubramanian et al. (2002) CurvesCurves
γφ,fP
E
Obtaining PrecipitationObtaining Precipitation
P
PETφ Where = Aridity Index
P
bγ Where = Soil Moisture Storage Index
γφ,fP
Q-P
max(S)max(E)bmax
max(S) = Soil Moisture Storage Capacity
PETP PET,
PETP P, max(E)
Estimating PETEstimating PET
• Temperature/Radiation-Based Methods:
1. Priestley-Taylor Method
2. Hargreaves Method
-Solar Radiation determined from:
a. VIC output
b. latitude, julian day, solar declination
• Combination Method (Penman):
- VIC output
Annual PET (1999)Annual PET (1999)(Hargreaves Method)(Hargreaves Method)
Dunne & Willmott (1996)Dunne & Willmott (1996)Soil Moisture CapacitySoil Moisture Capacity
1. Select Correction Domain and “Slope Bands”
2. Select Set of Basins that overlap with Correction Domain (must be gauged)
3. Determine “Actual” Basin Average Precipitation using Sankarasubramanian (2002)
4. Determine Scaling Ratios for each “Slope Band”
5. Apply the Scaling Ratios to an Existing Gridded Precipitation Dataset
Outline of StepsOutline of Steps
For Each Basin:
averaged over the same period of years
(as determined by the stream flow records)
Calculate Average Scaling RatiosCalculate Average Scaling Ratios
est
actave P
PR
Slope Bands in BasinSlope Bands in Basin
0
12
3
45
6
Create Scaling Ratios for each Create Scaling Ratios for each Slope Band (for each basin)Slope Band (for each basin)
est
actave P
PR Calculate Basin Average Scaling Ratio:
6
0sss
6
0ssave PrPR
Constraint:
6
0ss
654321
P
6P5P4P3P2PPdSolution:
1Rave
1,2,...60s for 1d
srs ,
Problem Set-Up:
Example for a Single BasinExample for a Single Basin
35P 80,P 100,P 260,P
1.4R
3210
ave
0.41-1.4
0.83580100260
3(35)2(80)100d
4.5r 4.0,r 3.0,r
2.5,r 2.0,r 1.5,r 1,r
654
3210
By Definition May Not Be Valid!
Also Works if Rave < 1
1. Separate Basins into Regions that are Climatologically/Hydrologically Similar
2. For each Slope Band (1 through 6), find an average scaling ratio for each region by weighting the scaling ratios for the basins within that region:
Weighting of Scaling RatiosWeighting of Scaling Ratios
...WWW
...rWrWrWr
s,3s,2s,1
s,3s,3s,2s,2s,1s,1aves,
i
i2,i1,is, Tcells
N
6
NW 3
Number of Valid Slope BandsNumber of Cells
within a Slope Band
1. Select Correction Domain and “Slope Bands”
2. Select Set of Basins that overlap with Correction Domain (must be gauged)
3. Determine “Actual” Basin Average Precipitation using Sankarasubramanian (2002)
4. Determine Scaling Ratios for each “Slope Band”
5. Apply the Scaling Ratios to an Existing Gridded Precipitation Dataset
Outline of StepsOutline of Steps
Application over North AmericaApplication over North America
1. Break the Continent into “Correction Regions”
2. Choose Streamflow Stations
3. Calculate Precipitation Scaling Ratios
4. Apply to Gridded Precipitation
Correction RegionsCorrection Regions
5
6
1
27
8
10
43
11
12
1314
9
1. Climate Classification
2. Basin Boundaries
3. Location of Streamflow Gauges
Koppen Climate ClassificationKoppen Climate Classification
Hydro1k Basin LevelsHydro1k Basin Levels
Level I
Level III
Level II
Level IV
1. Years of Operation
2. Drainage Area
3. Location4. Nesting5. Degree of
Management
Gauge Station Selection CriteriaGauge Station Selection Criteria
Porcupine River
Liard River
S. Sasketchewan River
Arkansas River
Missouri River
Santiago River
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Rave 0 1 2 3 4 5 6
Final ThoughtsFinal Thoughts
1. First application of this kind on a global scale
2. Problems will arise in regions where there are few or no streamflow gauge stations (Himalayas).
3. Sensitivity to delineation of “Correction Regions”
4. Much fine-tuning left to be done!